1

2

DUTY, IDENTITY, CREDIBILITY

Fake news and the ordinary citizen in

India

Lead author: Santanu Chakrabarti, Ph.D, Head of Audiences Research team, BBC World Service

With Lucile Stengel and Sapna Solanki (Senior Researchers, BBC World Service Audiences Research team).

And the teams at The Third Eye, and Synthesis.

A note on the authorship of this report:

This has been a truly collaborative project between the World

Service Audiences Research team and its agency partners. The

ideas, words and phrases in this report have come from many

places, including from discussions and meetings with the

various individuals named above, as well as in presentations,

emails, and conversations with all our partner agencies. In

particular, the lead author would like to credit our research

agency partners: one, The Third Eye, our qualitative research

partner, many of whose sentences - and insightful analysis-

have been used nearly unaltered in significant sections of the

report, and two, Synthesis, our data science partner, who came

up with sophisticated techniques - and easy to understand

explanations - for understanding the workings of a very

complex phenomenon. The lead author however takes full

responsibility for any lack of clarity in the report.1

This is a work of empirical evidence, not of opinions. Nonetheless,

any opinions that have slipped through in the report are solely those

of the lead author.

Nothing in this report is to be construed as indicating or reflecting

the corporate policy of the BBC, BBC World Service or any related

entity.

Google & Twitter are co-sponsors of this independent research

conducted by BBC and its partners. Google and Twitter had no input or

influence over the scope, research methodology, process or final report.

1 As indeed he does for the overuse of footnotes. Starting with this one.

3

Credits and Acknowledgments

The BBC World Service Audiences Research team

would like to thank:

Jamie Angus, Director, BBC World Service Group, for

commissioning the work

Marie Helly, Tess Colley, Marysia Novak, Alex Milner

and Joel Gunter for pushing us to think in ways we

werent accustomed to doing. And other colleagues for

discussions and suggestions

Trushar Barot, for making all of this happen

Catherine Blizzard, Director, World Service Marketing

& Audiences, for her unstinting support

Dejan Calovski, for reading the report and ensuring

that every now and then research- ese was replaced by

human language

The rest of the World Service Audience Insight team,

for picking up the slack without complaining

Suey Kweon, for her invaluable intellectual help

Anne Orwell, without whom nothing would move

And finally, our research respondents, for giving us so

much of their time. We hope that if any of them read

this report they will recognise much of the

phenomenon we describe in this report

Report contributors:

BBC researchers:

Santanu Chakrabarti, Ph.D

Lucile Stengel

Sapna Solanki

Research partners:

Ethnography - The Third Eye

Gitanjali Ghate

Jasmeet Kaur Srivastava

Jayadevan Ambat

Aamir Siddiqui

Arjun Surendra

Saloni Garg

Swapnil Sakpal

Big data - Synthesis

Ankit Kalkar

Harriet Robertson

Aakash Gupta

Report design

Mule Design

Lucile Stengel

Sapna Solanki

Contacts for this research:

Santanu Chakrabarti: santanu.chakrabarti@bbc.co.uk

Lucile Stengel: Lucile.stengel@bbc.co.uk

Sapna Solanki: Sapna.solanki@bbc.co.uk

http://thirdeye.cc/http://www.syn-the-sis.io/mailto:santanu.chakrabarti@bbc.co.ukmailto:Lucile.stengel@bbc.co.ukmailto:Sapna.solanki@bbc.co.uk

4

CONTENTS

Executive summary

Introduction

Part I. The conditions for sharing news

without verification

1. The lines between various kinds of news really blurry

2. Scepticism about the medias motivations and

intentions

3. The digital deluge and the move to a high frequency

information environment

4. Copying mechanisms for dealing with the deluge

5. The broken link between consumption and sharing

Part II. The motivations behind sharing

1. Sharing to verify (within the network)

2. Sharing as a civic duty

3. Interlude: Emerging political identities in India today

4. Sharing as an expression and projection of identity

Part III. What is fake news?

INTERLUDE: WHY ATTEMPTS TO CURTAIL

FAKE NEWS FALL ON DEAF EARS

Part IV. The effective narratives of fake

news messages

1. The fake news messages circulating amongst those

with a right leaning identity

INTERLUDE: THREE EMMERGING RIGHT WING

IDENTITIES IN INDIA

2. The fake news messages circulating amongst the

left

Part V. Fact-checking and verification:

Limited and specific

1. Two groups that assess fake news better than

others and the techniques they use

2. Verification and the very specific use of Google

3. TV and Newspapers: Still some credibility (even if

bought out)

4. V.4. The challenges of fact checking - on and off

the network

Part VI. The fake news ecosystem

Part VI. Conclusions

5

A note on the quotes from research respondents used here:

Most of the research interviews were done in the language the respondent was most familiar with, usually the local

language. Where respondents spoke in English, they have been left as is, without any grammatical or syntactical

corrections. Where they did not speak in English, the interviews were translated into English and have been used in this

report. We have not edited the translations or corrected grammatical errors in the translations, as we believe it gives a

better flavour of the original interviews. We are aware that out of context some of these English translations are likely to

look odd to the non-Indian observer; or they might give a misleading or negative sense of the linguistic ability of

respondents. When quotes or sections of quotes have been used they have not been edited, other than condensing

occasionally. When this has happened, it has been shown with ellipsis (

6

Executive summary

In all of the heated discussion about fake news around the world, the one thing that has remained

subdued is the voice of the ordinary citizen. In this project, therefore, we started with a central

question: why does the ordinary citizen spread fake news, without verification? And if ordinary citizens

are concerned about fake news, as multiple surveys seem to show, how have they changed their

behaviour in response to that concern? We were also interested in understanding better the type of

fake news spread not as stories/urls but as images and memes - anecdotally known to be the key

method of information dissemination in private WhatsApp and Facebook feeds. We were however also

keenly interested in starting to explore the question of whether or not there was a fake news ecosystem

in India on social media.

Using a combination of in-depth qualitative/ethnographic and big data techniques, we found that:

1. There are certain conditions that are necessary for the spread of fake news. These are:

The blurring of lines between all types of news

Scepticism about the motivations of the news media

The flood of digital information and the shift to a high frequency news consumption world

7

The coping mechanisms for dealing with onrushing digital information. These are: selective

consumption, preference for images, sender primacy, source agnosticism, nature of the forum,

and feel over think

The broken link between consumption and sharing

The audiences sharing tactics for WhatsApp and Facebook

2. The motivations behind sharing are rich and complex and need to be understood to establish why fake news is shared.

These are:

Sharing to verify within the networks

Sharing as a civic duty

Sharing for nation building

Sharing as an expression of ones socio-political identity.

We discovered that socio-political identity plays a key role in sharing of fake news, especially

for those on the right. While we see multiple distinct identities emerging within the right, they

are all bound by common narratives, but there is no real unified sense of a left identity in

India; instead there are micro identities (eg Tamil, Bengali, Dalit), and even within that a deeply

held socio-political identity can sometimes get reduced to the level of an issue for others within

the broad left.

8

3. This means that four types of narratives in fake news messages are particularly effective at passing by the critical filters of a wide swathe of the citizenry.

These are:

Hindu power and superiority

Preservation and revival

Progress and national pride

Personality and prowess (of PM Modi)

That is, validation of identity trumps verification of facts. We also discovered that similar tactics

were used, by right leaning and left leaning fake news messages, but the volume of right leaning

fake news messages was much more prominent in most respondents phones.

4. There is some use of fact checking, For example the use of Google or going to television, but this is limited and specific. But even the

few groups that do engage in this kind of verification behaviour are prone to sharing unverified false

information if it resonates with their identity

9

5. There is an emerging fake news ecosystem on Twitter, where we find fake news sources and amplifiers on the political right to be much more densely

interconnected and intermeshed.

On the network analysis map produced during this research, we see that many more of the

Twitter handles that have published fake news sit in the pro-BJP cluster, than in the anti-BJP

cluster. On Facebook, we also see signs of a polarised nation, with indications that those most

engaged with politics also seem to take the most interest in sources of fake news. We also find on

Facebook that legitimate news sources, and sources known to have published fake news, are

followed by audiences with distinct interests.

We conclude that for the reasons discovered, checking the flow of fake news especially in their

current predominant form of images and memes - is likely to be extremely challenging. We suggest that

all actors - platforms, media organisations, government, civil society- need to come together to tackle

the problem, since it is, in fact, as much a social problem as a technological problem.

But in this, ordinary citizens too need to take more responsibility about sharing things without

verification - and surely platform centric solutions to help them can be found without compromising

too much the essential characteristics of the platform. We also recommend that journalists investigate

further whether or not there is an organised ecosystem of fake news production and dissemination.

10

DUTY, IDENTITY, CREDIBILITY

Fake news and the ordinary citizen

in India

INTRODUCTION There was a time, what a time it was, it was a time of

innocence, a time of confidences2. In that ancient era, the

term fake news was used to refer to a particular form of

satire; and commentary of that time grappled with the

troubling notion that young people might be getting their

information not from actual legitimate journalistic outlets, but

from such satire3 4.

We refer of course to the discussions in the early 2010s

around The Daily Show and The Colbert Report, the American

TV shows that blurred the lines between hard news and

satire. Difficult as it is to imagine at this particular moment in

2 Simon & Garfunkel. (1968). Bookends. Bookends. Universal Music. LP. 3 Amarasingam, Amarnath,ed. ,The Stewart/ Colbert Effect: Essays on the

Real Impact of Fake News (Jefferson: McFarland, 2011) 4 Baym, Geoffrey, From Cronkite to Colbert: The Evolution of Broadcast News.

(Boulder: Paradigm Publishers , 2010)

time, this form of fake news was even called some sort of

corrective to, and substitute, for mainstream journalism.5

Today, it is fair to say, the term fake news carries few

positive associations. Today it is an inarguably negative term,

irrespective of who is using it, though broadly speaking all

users of the terms refer to misleading or false information.

The term has been used by journalists and researchers in

conjunction with words such as crisis, or even democratic

crisis6. As is well known, influential politicians around the

world have taken up the term to connote any news that is

critical of them and their achievements.

Fake News in the Indian media

In the course of this project itself, we have found that

coverage of fake news in the Indian media over the last three

odd years has grown by nearly 200%, partly driven by the

Cambridge Analytica exposes at the time of state elections. In

all there have been 47,5437 news articles online about fake

news between January 2015 and September 2018. English

language media were the first to start talking about fake news

and continue to cover it most often, with vernacular media are

starting to engage with the issue more of late. (See Figure 1

below) 5 McChesney, Robert W., Foreword to The Stewart/ Colbert Effect. By

Amarnath Amarsingam (Jefferson: McFarland, 2011). 6 See, for example, https://qz.com/india/1335161/indias-fake-news-crisis-

to-worsen-ahead-of-election-oxford-study/ and

https://www.bbc.co.uk/news/technology-44967650 7 Please see methodology appendix for details.

https://qz.com/india/1335161/indias-fake-news-crisis-to-worsen-ahead-of-election-oxford-study/https://qz.com/india/1335161/indias-fake-news-crisis-to-worsen-ahead-of-election-oxford-study/https://www.bbc.co.uk/news/technology-44967650

11

Fig. 1: Coverage of fake news by Indian media (Jan 2015-

Sep 2018)

In the recent past the fake news discussion has mostly

revolved around reports of often horrifying violence, with

WhatsApp often seen to be at the heart of the violence. The

media coverage has not quite been about the use of fake news

websites, or fake website news stories masquerading as real,

as has often been the case in the American context. Of late, the

fake news story in India has very largely centred around the

technology (ie WhatsApp/ Facebook) and the violence. While

the combination of technology and violence has naturally led

to some gripping headlines8, we find9 that the English

8 See for example, these stories: https://www.bbc.com/news/world-asia-india-

45140158; https://globalnews.ca/news/4333499/india-whatsapp-lynchings-

language media coverage of fake news in India has spanned

across a number of topics. Unsurprisingly, 46% of the

coverage is domestic, though 15% is about fake news in the

international context. One of the most important points to

note here, though, is that reporting on fake news related to

scams and scares constitutes just 0.7% of the coverage. (This,

as we will see later, is quite the inverse of the topics audiences

are sharing on their WhatsApp feeds). Interestingly, just 9% of

the coverage is about solutions to the fake news problem,

indicating the complexity of the situation.

Gone missing: The ordinary citizen

A lot of the media commentary - and emerging research - on

the phenomenon has focused on the actors responsible for

creating fake news (e.g. Macedonian teenagers from the town

of Veles10; or suspected Russian state actors11), the platforms

thought to play a central role in the spread of fake news (e.g.

falsehoods spread faster on Twitter than does truth12, the

child-kidnappers-fake-news/; https://www.wired.com/story/how-facebooks-

rise-fueled-chaos-and-confusion-in-myanmar/ 9 See Methodology appendix for more details 10 One of the best of these stories is this one:

https://www.wired.com/2017/02/veles-macedonia-fake-news/ 11 See for example:

https://www.theguardian.com/technology/2018/jul/31/facebook-russia-

election-midterms-meddling; 12 Vosoughi, S., Roy, D. , Aral, S. The spread of true and false news online (Science,

2018) 1146 -1151. 13 See https://www.nytimes.com/2018/10/17/opinion/brazil-election-fake-

news-whatsapp.html?module=inline

https://www.bbc.com/news/world-asia-india-45140158https://www.bbc.com/news/world-asia-india-45140158https://globalnews.ca/news/4333499/india-whatsapp-lynchings-child-kidnappers-fake-news/https://globalnews.ca/news/4333499/india-whatsapp-lynchings-child-kidnappers-fake-news/https://www.wired.com/2017/02/veles-macedonia-fake-news/https://www.theguardian.com/technology/2018/jul/31/facebook-russia-election-midterms-meddlinghttps://www.theguardian.com/technology/2018/jul/31/facebook-russia-election-midterms-meddlinghttps://www.nytimes.com/2018/10/17/opinion/brazil-election-fake-news-whatsapp.html?module=inlinehttps://www.nytimes.com/2018/10/17/opinion/brazil-election-fake-news-whatsapp.html?module=inline

12

possible use of WhatsApp to poison Brazilian politics13) or

indeed the victims of fake news - from individuals to

communities to even entire democracies. A whole host of

academics from a whole host of disciplines - including

economics, computer science, have sought to understand

various aspects of the phenomenon, from the consequences of

digital information overload to the economics of fake news

production.

A lot of the research in the area has been focused on the

technology or platform and used technology in the analysis14.

In all of the frenetic and frantic research and commentary,

there is one thing that has gone underexplored: the voice of

the ordinary citizen - and indeed, the responsibilities of the

ordinary citizen. Where the ordinary citizen does appear,

especially in the media, s/he is sometimes inadvertently

portrayed as dupe of malicious actors, or heavily influenced

by social media/chat app messages to the extent of

committing acts of egregious harm.15 Despite the injunctions

of researchers like Alice Marwick, who call for a

sociotechnical approach to understanding the fake news

phenomenon, and Wardle & Derakshan16, who draw upon the

work of James Carey and urge researchers to understand

14 See for example: Qiu, Xiaoyan, Oliveira, Diego F. M., Shirazi, Alireza Sahami,

Flammini, Alessandro Menczer, Filippo, Limited individual attention and online

virality of low-quality information (Nature Human Behavior, 2017), 1-132. 15 For example, the headline here says How WhatsApp helped turn an Indian

village into a lynch mob: https://www.bbc.com/news/world-asia-india-

44856910 16 Wardle, Claire, Derakhshan, Hossein . Information Disorder (Strasbourg,

France: Council of Europe, 2017)

better the ritualistic function of communication, some bits of

the academic research and a lot of the journalism so far have

inadvertently diminished the agency of ordinary human

beings, focusing more on the technology of the

communication. Even Wardle & Derakshan, who do think

about communication as culture, recommend action by

technology companies, national governments, media

organisations, civil society, education ministries, and funding

bodies - but dont have anything to say about what the

responsibility of the ordinary citizen is in addressing the

problem of fake news.

The occasional study has centred the publics voice/role on

the phenomenon of fake news. A study commissioned by the

BBCs commercial news arm, Global News Ltd., found that

79% of the public in 6 countries of the APAC region were very

concerned about the spread of fake news.17 A late 2017 global

poll in 18 countries conducted by Globescan for the BBC

World Service found that 79% of the respondents globally

were concerned about whats fake and whats real on the

Internet.18 In Kenya, 87% of the respondents in a study by

Portland reported that they had seen deliberately false

news.19 All of these studies shed some light upon the

phenomenon as experienced by the public. However, they

answer more of the what questions than the why or the how

17 BBC Global News Limited, The value of news (BBC Global News, 2017) 18 See: https://www.bbc.com/news/technology-41319683 19 Available here: https://portland-communications.com/pdf/The-Reality-of-

Fake-News-in-Kenya.pdf; the study did not go into the question of how

respondents knew the content was deliberately fake.

https://www.bbc.com/news/world-asia-india-44856910https://www.bbc.com/news/world-asia-india-44856910https://portland-communications.com/pdf/The-Reality-of-Fake-News-in-Kenya.pdfhttps://portland-communications.com/pdf/The-Reality-of-Fake-News-in-Kenya.pdf

13

questions about citizens motivations and behaviours20.But

because the research terrain is relatively new, these studies

lead to more questions than answers. Above all, though, the

main question that these studies throw up is this: what

exactly is this concern that people seem to be expressing?

And if indeed they are so concerned, how have they changed

their behaviour in response to that concern21? It is well-nigh

impossible to answer these questions by simply tracking their

behaviour on technology platforms or even by asking them a

set of questions in a large scale quantitative survey. It is by

immersing ourselves to some extent in the lives of these

ordinary citizens to some extent can we start to understand

how social and cultural forces, and their own desires and

aspirations, come together to influence the role they play in

the spread and growth of fake news.

Whither India?

In addition to this people sized hole at the centre of many of

the research projects and papers, much of the published or

20 This is as it should be - quantitative surveys are better at answering certain

questions than others. However, it must be said that online quantitative studies

are increasingly easier and quicker to do than ever before, and many

journalists find it easier to report on such studies than others. Consequently,

quantitative surveys are often being used for purposes they are not designed

to fulfil. 21 We are assuming that concern usually leads to some kind of change, first

in the individuals internal state of being and then in their external actions.

We recognize though that there is a body of scientific evidence that argues

this is not inevitable and that existing human behaviour, say, for example,

humanitys response to the threat of climate change, is ample evidence of

the big gap that exists between attitudes and behaviour. (Which also leads

us to the depressing conclusion that the movie Wall-Es depiction of future

humans might be extremely prescient).

publicly available research has been centred in the USA, and

to a certain extent, Europe. While there has been reporting on

fake news and its consequences from around the world and

extensive coverage of the issue in local media in many

countries of the world, this has not yet been accompanied by a

similar volume of published research in those countries. In

India, especially, much of the discussion seems to have been

conducted in the media, and not so much in the academy or

within think tanks.22 Of course, the research cycle moves

slowly while the news cycle moves at the speed of light, but

the lack is glaring; and the first draft of history, which is what

journalists write, should not become the final verdict on the

phenomenon.

And whats going on inside private WhatsApp and

Facebook feeds?

In addition, it is well-nigh impossible for researchers to use

algorithmic/computerized/automated techniques to

investigate audience behaviour within encrypted private

networks (eg Facebook, WhatsApp) - and this is as it should be

to ensure peoples privacy.23 As a result, the picture we have

about how people are sharing information, especially fake

news, is from the outside, so to speak, assessed primarily

22

There have been research studies looking into government/ state disinformation which have included India, for example:

http://comprop.oii.ox.ac.uk/wp-

content/uploads/sites/93/2018/07/ct2018.pdf 23 In this paper we do not engage with the debate on fundamental issues of

privacy and data collection on technology platforms.

http://comprop.oii.ox.ac.uk/wp-content/uploads/sites/93/2018/07/ct2018.pdfhttp://comprop.oii.ox.ac.uk/wp-content/uploads/sites/93/2018/07/ct2018.pdf

14

from the metrics that are available within the analytics

systems of the platforms24. This picture is especially fuzzy

when it comes to the WhatsApp and Facebook. Not that much

is known25 of the content of what is actually in the feeds/

groups of users. Nor is that much known about the actual

strategies and tactics people adopt to share messages on

WhatsApp. WhatsApp, for example, has been put at the heart

of Indian coverage of fake news and violence, but it has not

been fully explained, which innate characteristics of

WhatsApp have made it so central to the debate. Or, talking

about Facebook, or Twitter, or any other platform for that

matter, how exactly and for what purposes - ordinary

citizens are using these platforms - and how that matters in

the context of fake news. Not that much work has been done

either in categorising the messages within these networks by

content, even though there have been proposals to categorise

these by intent 26.

24 Here we are referring mainly to Facebook and Twitter. WhatsApp lacks even

rudimentary analytics systems, or at least does not make it available widely -

this is a feature, not a bug, from the perspectives of its founds, a consequence

of their commitment to user privacy and lack of interest in advertising. See for

example, https://www.theguardian.com/technology/2018/apr/30/jan-koum-

whatsapp-co-founder-quits-facebook. 25 At least not to the external world; Facebook, for example, will know about

the material flagged for moderation. 26 Wardle, Claire, Derakhshan, Hossein . Information Disorder (Strasbourg,

France: Council of Europe, 2017)

What do we mean when we use the term fake news in this

report?

As we have touched upon earlier, there are multiple uses - and

abuses - of the term fake news. While a precise definition of

the object of inquiry is critical to high quality research, one of

the objectives of this project was to understand how ordinary

citizens defined the term, if at all. That being said, Wardles

categorisation27 of mis and disinformation into seven broad

categories of satire/parody, misleading content, imposter

content, fabricated content, false connection, false context,

and manipulated content, was certainly a starting point.

However, classifying fake news by the intent into mis, dis

and mal-information seems problematic, because judging

intent from outcome is - as journalists well know28- no easy

task. Also, including satire/ parody in the bucket of mis/dis-

information because it has the potential to fool29 sat

uncomfortably with us: not just because satire has been

historically a weapon of the weak against the powerful but

also because we suspected that most people for the most part

did have the ability not to be fooled by satire.

In this report, we will use the term fake news30. Our rationale

for doing so, as opposed to the variety of other terms proposed

27 Ibid. 28 See, this: https://www.nytimes.com/2018/06/25/reader-center/donald-

trump-lies-falsehoods.html 29 Wardle, Claire, Derakhshan, Hossein . Information Disorder (Strasbourg,

France: Council of Europe, 2017) p17 30 A number of prominent social scientists observed in an article in Science, that

just because a terms has been weaponised should not mean we do not use it.

See: http://science.sciencemag.org/content/359/6380/1094

https://www.theguardian.com/technology/2018/apr/30/jan-koum-whatsapp-co-founder-quits-facebookhttps://www.theguardian.com/technology/2018/apr/30/jan-koum-whatsapp-co-founder-quits-facebookhttps://www.nytimes.com/2018/06/25/reader-center/donald-trump-lies-falsehoods.htmlhttps://www.nytimes.com/2018/06/25/reader-center/donald-trump-lies-falsehoods.htmlhttp://science.sciencemag.org/content/359/6380/1094

15

such as mis-information/mal-information/disinformation31,

junk news 32, or even propaganda, is in line with many of the

arguments of Marwick33. However, while Marwick, borrowing

from Caroline Jack, prefers to use the term problematic

information, we persist in using the term fake news for the

simple reason that this project starts with the emic

perspective34, even if it does not end there as we will see later,

that the term that the ordinary citizen uses to refer to the

entire spectrum of incorrect or misleading information (and

more) is - for good or for bad - fake news. It is mainly for this

reason that we use the term fake news35 in this report- and not

just as a heuristic. Note, though, that while we were keenly

interested in understanding how citizens perceived fake news,

we did not use their definitions in the analysis of news stories

or the content of their private networks. We have used Indian

fact-checking sites such as altnews and boomlive to assess the

truth claims of news articles called out as fake. Researchers on

this project have also assessed the truth claims of various

pieces of private network content. As a starting point, then,

our definition of fake news was this: information, however

31 Wardle, Claire, Derakhshan, Hossein . Information Disorder (Strasbourg,

France: Council of Europe, 2017) p17 32 Bradshaw, Samantha, Howard, Phillip N.,Why does junk news spread so quickly

across social media ? Algorithms, advertising and exposure in public life (Oxford

Internet Institute, 2018) 33 Marwick, Alice E.,Why do people share fake news: a sociotechnical model of

media effect (Geo. L. Tech.Rev.474, 2018). 34 For a good- but very specialist discussion- see Harris, Marvin, History and

Significance of the Emic/Etic Distinction (Annual Review of Anthropology, 1976)

329-50. 35 We report relief at this point to be able to drop the quotation marks from

fake news!

conveyed, in whichever format, on whatever platform, which is

not fully supported by factual evidence.36 That is, our starting

definition of fake news certainly goes much beyond the news

stories on websites, located somewhere on the internet,

available by using an url, and shareable on social media

platforms using that same url. We deliberately include all

formats of information sharing, primarily because we

anecdotally know that the dominant form of information

sharing on private networks in India is not through news

stories as defined above, but very much in the form of images,

pictures, memes, etc.

Approaching this project: Ordinary citizen, sharing, and

verification

Our starting observation for this project was the simple

observation that there are a few basic conditions that are

required for fake news to spread through networks. It

certainly needs someone to create the fake news, and it

certainly needs platforms and technologies which enable

them to spread. But it also needs one critical element:

ordinary citizens to share the fake news in their networks.

And it needs these things to be spread on without verification.

For us, then, an understanding of the fake news phenomenon

is simply incomplete without understanding this key

question: why does the ordinary citizen share fake news without

36 We note here that the issue of what constitutes evidence can be debated, of

course. In India, for example, debates around history in India often centre

around what constitutes evidence. For the purposes of this project, we follow

the standards of factual evidence that are regarded as normative in the

academy and can be arrived at by a process of inductive or deductive reasoning.

16

verification? The simplest and therefore simplistic37 answer is

that they dont care about facts and accuracy or even truth.

Equally reductive is any sense that the majority of ordinary

citizens spreading fake news are either malicious or

duplicitous, or conversely dupes of malicious state actors.

None of these assumptions and explanations seemed

sufficient to us as researchers, leave alone the necessity to

explain the phenomenon of fake news.

To be absolutely clear, our focus in this project is neither state

actors nor government propaganda. In this project, we aim to

shift the focus to the actions of ordinary citizens. We aim to

understand the fake news phenomenon as a socio-cultural

and socio-political phenomenon enabled by technology rather

than as a purely technological phenomenon.

In this project then, we draw upon the British Cultural Studies

research tradition to understand the usage38, but in particular,

the sharing of fake news by ordinary citizens, both within

encrypted Facebook networks and within WhatsApp. And in

particular, we aim to understand if and why sharing happens

without verification. We attempt to situate this sharing

activity of citizens in their specific sociocultural contexts. In

this project, therefore, we start with people and the nature,

content, and structure of the messages they are sharing, but

37 And more common in more circles of discussion than anyone would care

to admit 38 Without arguing that media effects are minimal. In fact, one of the central

objectives of this project is to assess the effects of peoples networks on

their sharing behaviour.

we also look deeply at the way they are using platforms

(Facebook and WhatsApp, mainly). We pay attention to the

amplifiers39 and influencers of fake news messages as much as

we do to the platforms.

Fig. 2: How we are studying fake news in this project

39 We describe what we mean as amplifiers in the last section of the report

17

This project is exploratory in nature, and aims to approach

something close to a thick description40 of the phenomenon

of fake news. However, while being exploratory, it is also

conclusive in many of its findings, given the breadth and

depth of the fieldwork. We do hope this report will be read

widely by anyone interested in fake news or for that matter

India, but the principal audience we had in mind while writing

this report are academic researchers. We want this report to

spur exploration and research across multiple fields of inquiry,

and as such we have borrowed magpie like from multiple

disciplines41 in the writing of this report.

About the methodology and research process

We set out in this project to try and answer the question of

why ordinary citizens spread fake news without verification -

a little understood part of the fake news equation. The project

was extremely quick turnaround, from starting to final report

publication was to be completed in three months flat. This

meant that we would not be able to address all of the

questions we would like to answer and we would also not be

able to use all of the methodological tools at our disposal.

40 See Geertz, Clifford, The interpretation of cultures (New York: Basic Books, Inc,

Publishers,1973) 41 The lead author takes full responsibility for this cavalier approach to

disciplinary boundaries, and drawing upon literature he is only passingly

familiar with.

Key to deciding on approach was the fact that the fake news

phenomenon is very new indeed, and not yet that well

understood. When a phenomenon is new or not very well

understood, qualitative research techniques are especially

useful. These techniques - in this case, in-depth interviews

and up-close observation of sharing behaviours - allowed us to

explore fake news with nuance, richness and depth. And

because we wanted to know what was spreading in encrypted

private networks like Whatsapp, semi ethnographic

approaches in this case, visiting people at home - was

essential. This allowed us to understand the individual in full,

and establish how their histories and backgrounds had

brought them to the present point; and how they were

contending with societal and cultural forces surrounding

them.

We debated using large sample quantitative surveys42 but we

came to believe that the survey methodology should follow

the establishment of a conceptual framework and intellectual

scaffolding. Our interest was in exploring audience

psychology in-depth to start with, in particular to understand

what citizens meant by the term fake news, so we decided

against using survey methods for this project.

Fieldwork and analysis were carried out in six overlapping

stages (see Methodology Appendix for more detail on each

stage of the methodology):

42 Not least because we work in a journalistic organisation and have

observed that journalists feel the most comfortable reporting on research

whose findings are conveyed in charts and graphs!

18

1. Respondents were recruited and consent forms signed.

Respondents came from a mix of social, political, age,

gender and economic backgrounds. They were first

asked to share with researchers what they found

interesting in their WhatsApp and Facebook feeds and

were sharing within their networks. Researchers were

very careful to not use the term fake news at this

stage-because one of the key objects of the inquiry was

to assess whether or not respondents were able to

detect what was fake and what was not; and what they

labelled as fake.

2. Post seven days of such sharing, in-depth, in home

interviews were held, where researchers had detailed

conversations touching on multiple aspects of their

lives from childhood to adulthood, their influences,

their idols, their likes and dislikes, their reaction to

their changing environments, their news consumption,

their usage of digital platforms, their social, cultural,

and political perspectives and their sharing activity,

eventually arriving at the topic of fake news. During

the course of - and again at the end of - the sessions,

respondents were asked to show researchers the

contents of their WhatsApp and Facebook feeds, and

discussions were held on what they would share, would

not share, and why. Also, they were shown known fake

news messages and asked whether or not they found

these credible and why. Respondents were finally

asked to share with researchers. Their interviews were

then analysed using a grounded theory approach43 -

given that this was a new phenomenon being explored-

and the results married to what was being found in the

analysis of the messages.

3. The messages shared by participants were in parallel

analysed by semioticians on the one hand, and data

scientists using machine learning methods on the other

- for tone, content, style and structure.

4. Data science approaches were then used to retrieve and

analyse the media coverage of fake news in India, in

English and vernacular languages.

5. After a list of fake news sources was generated in

stages 3 and 4, the Twitter network of fake news

sources and amplifiers in India was mapped44 , and

clustering analysis was conducted to understand

agents sharing similar connections.

6. Publicly available Facebook advertising data was used

to understand strength of affinity between audiences of

fake news sources and audiences of legitimate news

sources, and build a network map of these various

communities.

That is, the project ended up using multiple methodologies

eventually:

o In-depth in home qualitative and semi-

ethnographic approaches: 120+ hours of in

depth interviews at home across 10 cities and 40

43 See methodology appendix for more details 44 See methodology appendix for more details

19

individuals in India. The cities in which the

interviews were conducted were a mix of metros

and smaller towns, spanning the north, south,

east, and west of the country: Mumbai, Delhi,

Kolkata, Chennai, Rajkot, Vijayawada, Raipur,

Udaipur, Amritsar, Varanasi. The interviews

were conducted in the language the respondent

was most comfortable in- in most cases the local

language. The interviews were transcribed and

translated. Original recordings, photographs and

videos taken inside the respondents homes

were used for analysis, alongside English

language translations of the conversations.

o Auto ethnography: Collection of a corpus of fake

news messages.

o Semiotic analysis: Understanding signs,

symbols, and structures of fake news messages.

o Big data/ network analysis: Across 16,000

Twitter profiles (370,999 relationships); 3,200

Facebook pages & interests

o News scan and topic modelling: Media scans

from last two years of news about fake news, in

English and in local languages. 47,543 in total.

The findings from all of these stages- plus learnings from desk

research/ review of existing literature- was brought together

by the research teams to create initial presentations, finally

followed by this report that you are reading.

For taking the time to read this report, thank you.

And now, heres what we found.

20

The findings

21

In the next few45 pages we will tell you the story of fake news

and ordinary citizens in India as we see it from our analysis of

the data. We will tell you this story in six main parts.

These are:

1. The breeding ground: In which we outline what our analysis suggests are the necessary conditions for the spread

of fake news. We outline six conditions from the news

environment to psychological attributes to citizens sharing

strategies on the sharing platforms.

2. The motivations behind sharing: In which we try to understand why people share. In an interlude, we outline our

observations on the socio-political identities we see

crystallising in India today- as we find this quite central to

understanding sharing behaviour by citizens

3. What is fake news: In which we address this question from the perspective of the citizens

4. The narratives of effective fake news messages: In which

we understand the narratives that effectively short-circuit

critical thinking (And we look at fake news messages of the

left and right) effective

45 Ok. Its not exactly few.

5. Fact-checking and verification: In which we address why verifying behaviour is limited and specific

6. The fake news ecosystem on social media (Twitter and Facebook): In which we try to understand if there is a fake

news ecosystem, and see how sources known to have shared

fake news, are interacting with political and media actors.

7. Conclusions and suggestions for further research. Does what it says on the tin

22

The breeding ground The conditions created for sharing news without verification 1

23

I. The breeding ground: the

conditions created for

sharing news without

verification

I.1. The lines between various kinds of news have

blurred

When discussing fake news, both parts of the term are

equally important. So first, we must understand how people

understand the term news, before we move onto issues of

fake and genuine.

As a term, news itself has always had more than one

meaning: on the one hand, news is what you got in your

newspapers and televisions and radio sets; on the other hand,

news is also information about you, your family, and others

important to you. In the realm of institutional news providers

- and researchers of media - too, there has always been a

further demarcation between hard news and soft news 46 as

there has been one between news and features. Many of

46 See, for example, Reinemann, Carsten , Stanyer, James, Scherr,

Sebastian, Legnante, Guido. Hard and soft news: A review of concepts,

operationalizations and key findings. (Journalism, 2012) 221-240

these demarcations originated from the world of print

newspapers, and were carried over into TV news when it first

started. But the emergence of Facebook as a key platform for

news and the centrality of its newsfeed- established a

forum where not only hard and soft news and news and

features blended together; it also created a forum where

news about your nieces birthday and news about a

dictators latest autocratic actions merrily intermingle. A third

demarcation is of most importance for our purposes - the

demarcation between fact (i.e. just the pure reporting) and

opinion, with the category of analysis lying somewhere in

between. This has been a key feature of traditional journalism

with its routines, structures and adherence to the norms of

objectivity and/or impartiality.47 It is this distinction between

fact/reporting and opinion that seems to have been almost

completely decimated by digital news sources, especially by

social platforms, such as Facebook.

This, at the level of the ordinary citizen has had certain very

important consequences. In India, we see amongst our

respondents an upending of the traditional divide between

news and opinion. For the respondents, news is primarily

47 These are related but different terms, with objectivity more commonly

used in the American journalism context. For example, see the BBCs editorial

guidelines ( https://www.bbc.co.uk/editorialguidelines/guidelines/bbc-

editorial-values/editorial-values) for how it thinks about the idea of

impartiality, and Tuchmans classic sociological investigation of the workings

of objectivity in an American newsroom: Gaye, Tuchman. "Objectivity as a

Strategic Ritual: An Examination of Newsmen's Notions of Objectivity."

(American Journal of Sociology , 1972) 660-679

https://www.bbc.co.uk/editorialguidelines/guidelines/bbc-editorial-values/editorial-valueshttps://www.bbc.co.uk/editorialguidelines/guidelines/bbc-editorial-values/editorial-values

24

divided into two categories to start with. First, citizens think

in terms of information, which is related to the fact that the

headline conveys - and mostly that. And then, there is news

which is the actual content. It is acknowledged and largely

desired that there be a layering of opinion - or at least a stance

- on the facts. It is then this layering in of opinion on to facts

that makes it news.

Crucially, news is now considered to be as much about how it

makes me feel as about what it tells me. Human interest

stories, or softer news in general is considered to be a core

part of news, while political and policy reportage is expected

to not just be dry and analytical, but to express some emotion.

In other words, people expect news to not just inform but to

entertain.

This is not an entirely new phenomenon. This collapsing of

boundaries between various types of news - albeit many of the

boundaries said more about news organisation norms than

news consumer needs- predates the rise of digital and social

media. Media scholars have been expressing anxiety about the

blurring of news and entertainment for a while now,

lamenting the rise of global infotainment and the

Bollywoodization of news.48 It will take us outside of the

purview of this report, so we will not develop the idea at

length: but it seems to us that one of the ways to understand

this development is to see it as the mainstreaming of the

48 Thussu, Daya Kissan. News as Entertainment: The Rise of Global

Infotainment. (London: Sage Publications, 2007)

codes and norms of Indian vernacular news media - which

have traditionally been quite different from those of the

English language media in India49.

What is new though, is that with the definition of news

becoming expansive and all encompassing, we find that

anything of importance to the citizen is now considered news.

It then also stands to reason that they are happy receiving

information from just about anyone and not just a handful of

news organisations with rigorous journalistic practices,

trusted brand images, and legacies of accuracy. Even more

importantly for our purposes, we find that people dont

differentiate or more accurately find it too hard or too

resource intensive50 - to differentiate between various sources

of news (in the broad sense outlined above). Social media, with

its low barriers to entry provides innumerable sources of

information - and the distinctions between them are flattened

in the minds of the users. In fact, as we detail later, the

identity of the source is not at all central to the question of

consumption or sharing of information. But this is - at least

partly - a function of the sheer volume of digital information

that the average Indian user of WhatsApp and Facebook now

receives in their feeds every day.

49 See, for a discussion, Rajagopal, Arvind. Politics After Television: Hindu

Nationalism and the Reshaping of the Public in India. (Cambridge,UK:

Cambridge University Press., 2001) 50 Here, when we say resource we primarily mean time and cognitive

resources. See later sections where we discuss this in detail.

25

I.2. Scepticism about the medias motivations and

intentions

Amongst our respondents, we see a loss of trust about

the intentions of the news media, accompanied with a longing

for a past era when all was well and journalism was of the

highest standards. There is widespread belief that TV- and to

some extent, print - are motivated by profit more than

anything else; if anything, for todays citizens, the agenda of

the media isnt even so much hidden. The media is certainly

felt to be more biased and partisan today than it was before,

and the credibility of journalism per se is not dramatically

high. This attitude, though, comes from exposure to satellite

TV stations, particularly in the local languages, which are

known for their shrill and sensationalistic tone and hyperbolic

presentation style. Those whose media mix includes

newspapers are not that negative about journalism. (And

Doordarshan,51 considered staid and boring by many, does

have some amount of residual affection - in part because staid

and boring stands out in marked contrast to sensationalistic

and hyperbolic!) But overall, the sense is that the media has its

own agenda, stemming from its affiliation, with either

business, or political interests. There is a caveat emptor

attitude when it comes to the output of Indian media, especially

that of TV news channels.

51 The Indian state broadcaster, now a network of channels. Still consumed

widely, despite the number of private players in the market. See:

http://www.indiantelevision.com/television/tv-channels/viewership/urban-

areas-dominate-dd-viewership-zapr-180225

M: News is coming from your fathers time, so the news

which you get at those days like the way of telling, the

topics, in all things comparing it with the news now what

it is, is there any difference?

R: At that time, what the reality in news was there it has come down compare to now. At that

time, Doordarshan mean like news telling etc. and it was real news.

(Male, 25, Vijayawada)

Like if you see Doordarshan, the host is a simple one and just read the news but the private news channel they make the small news into a breaking news. To increase their TRP, they will hype that small news to a breaking channel. (Female, 25, Udaipur)

If we take TV news, each channel is biased with some party, so there is no one neutral. In

newspaper also, each paper is biased with some party. Only the social websites have common

news because it is uploaded by a common people only, so I feel that, this is only unbiased news.

(Male, 26, Chennai)

http://www.indiantelevision.com/television/tv-channels/viewership/urban-areas-dominate-dd-viewership-zapr-180225http://www.indiantelevision.com/television/tv-channels/viewership/urban-areas-dominate-dd-viewership-zapr-180225

26

Note here the last line of the last comment. This will be of

importance later in the story.

Fake news was not there as on Doordarshan and

CNN and BBC would put paid content in biased

view, but they won't put fake and untrue things.

I.3. The digital deluge- and the move to a high

frequency information environment

Indians today are having to deal with a huge volume of

digital information in their WhatsApp and Facebook feeds.

This is partly a result of the recent and dramatic drops in the

cost of data as well as the costs of smart phones and

subsidised feature phones.52 Since data costs are no longer a

key constraining factor in the use of digital networks, unlike

their peers in Kenya and Nigeria, people simply do not have to

52 See for example, Majumdar, Romita. Two years of Jio: How free calls,

data catalysed India's digital revolution, Business Standard. 6 September

2018. https://www.business-standard.com/article/companies/two-years-

of-jio-how-free-calls-data-catalysed-india-s-digital-revolution-

118090501169_1.html (accessed November 3, 2018).

think twice about sharing digital media, irrespective of the

nature of said media.

As a result, we find our respondents inundated with messages

on WhatsApp and Facebook. There is a near constant flurry of

notifications and forwards throughout the day on their

phones - encompassing from news organisations updates to a

mindboggling variety of social messages (for example,

inspiring quotes and good morning forwards, the latter of

which seems to be a peculiarly Indian phenomenon, even the

subject of discussion in the international media).53

News providers - and there are tens of thousands of them in

India54- do not make it any easier, by sending regular, even

incessant, notifications to phones.

I am attached to social media, and I have an app of ABP news and it updates me in every 2-4 minutes about job and at present it is coming, and it may have a lot of notifications of news. (Male, 34, Amritsar)

53 See, for example, Purnell, Newley. The Internet Is Filling Up Because

Indians Are Sending Millions of Good Morning! Texts. The Wall Street

Journal. 22 January 2018. https://www.wsj.com/articles/the-internet-is-

filling-up-because-indians-are-sending-millions-of-good-morning-texts-

1516640068 (accessed November 4, 2018)

54 Source for saying the number of news providers available

https://www.wsj.com/articles/the-internet-is-filling-up-because-indians-are-sending-millions-of-good-morning-texts-1516640068https://www.wsj.com/articles/the-internet-is-filling-up-because-indians-are-sending-millions-of-good-morning-texts-1516640068https://www.wsj.com/articles/the-internet-is-filling-up-because-indians-are-sending-millions-of-good-morning-texts-1516640068

27

The default behaviour for our respondents at the moment

seems to be to have notifications on rather than off- and we

believe this behaviour is quite widespread, for many

respondents, when asked how they come to know about a

news event, say that its because of notifications. The rise of

digital as the central means of acquiring information has

effectively transformed news behaviour from a high time spent,

low frequency behavioural pattern to a high frequency, low time

spent behavioural pattern. In other words, people have many

more moments of news consumption than ever before, but

each of these moments is now of really short duration. In

India, citizens actively seem to be privileging breadth of

information over depth, partly as a response to the digital

deluge. This, we find, is necessitating the adoption of coping

mechanisms which then end up facilitating the spread of

fake news.

I.4. Coping mechanisms for dealing with the

deluge

To be absolutely clear, Indians at this moment are not

themselves articulating any kind of anxiety about dealing

with the flood of information in their phones. If anything, they

only see the positives of social media.

Because contacting has become easier due to social media, people are easily connected on WhatsApp, Facebook, there you relate to each other, you get to know the birthdates of people, you keep on getting texts, notifications about people. Earlier this was not there, earlier people needed to meet each other, talk face to face. But now things have changed, people talk with each other on WhatsApp, so that change has happened. Connection has become easier now (Respondent, Male, 63, Delhi).

In India, the need to stay updated (and that is exactly the

term that respondents use) overrides any anxiety about

needing to cope with the constant stream of digital signals

asking for attention. Over the years, in multiple research

projects, including this one, we have repeatedly picked up a

desire from Indian citizens to somehow stay abreast of

everything that is going on. This is especially pronounced

among the young- the reasons for this are multifaceted, from

not wanting to come across as uninformed in a peer group, to

the pressures coming from the countrys competitive entrance

exams, many of which test candidates on general knowledge.

Our knowledge will be increased if we get updated with the news, so if tomorrow we go in any interview or we gets attached to any website,

28

or communicate with any customer, then we give them information. (Male, 34, Amritsar)

In effect, navigating WhatsApp and Facebook is now part of

everyday life and people are doing it without consciously

thinking about how they are doing so. That does not mean

though that it is easy to do. WhatsApp and Facebook - which

we will from this point refer to collectively as digital sharing

platforms- are quite likely leading to a situation of

information overload.55 For example, refer to the quote earlier

from the male respondent in Delhi; he says you keep on

getting texts, notifications about people. The telling phrase

there is: keep on, and we hear similar use of language to talk

about information flows in these networks, but while it

sometimes flows over into irritation, a tone of anxiety or

concern about the difficulty of managing these flows does not

really come through.

55 Information overload is not the most well-defined of terms, despite

being used in diverse fields such as cognitive science, business, and

technology. Clay Shirky argues that the problem is not information

overload, because we have always been dealing with information overload,

but filter failure there are no economic incentives for producers of digital

content to filter for quality before publication. See Juskalian, Russ.

Interview with Clay Shirky Part I. Columbia Journalism Review. 19

December 2008

https://archives.cjr.org/overload/interview_with_clay_shirky_par.php?page

=all (accessed November 4, 2018).

... You dont want, but still, the posts are coming through, right? The people keep sharing it, its very irritatingConstantly, you will keep getting. There are groups in that people have their political affiliations. The regular ones have come, it's Okay. Political things keep coming. It's very irritating. (Male, 38, Mumbai)

But while people are not articulating their anxieties, we find

that they are certainly adopting a slew of tactics and

stratagems to cope with their digital feeds. It seems, though,

that as Qiu et al56 have suggested when reviewing the

literature on cognition in computer mediated/ digital

environments, that paradoxically, our behavioural

mechanisms to cope with information overload may make

online information markets less meritocratic57 and diverse,

increasing the spread of misinformation and making us

vulnerable to manipulation.

56 Xiaoyan Qiu, Diego F.M. Oliviera, Alireza Sahami Shirazi, Alessandro

Flammini, Filippo Menczer. Limited individual attention and online virality

of low-quality information . Nature Human Behaviour, 2017, 4

57 The approach here is based in the field of economics, so merit and quality

are not limited to ideas of the accuracy of a piece of information.

29

The strategies that are being adopted by people include:

I.4.1 Selective consumption58: This operates in two ways. First,

given the volume of messages in their feeds, a significant

proportion of the messages received in their feeds, especially

WhatsApp feeds, are simply not opened or consumed. Second,

messages are often part consumed before they are forwarded

on. This part consumption could be based simply on the

headline without actually consuming the content59 This can

sometimes lead to situations where messages are forwarded

on, whose content is not even subscribed to by the sender, but

they are sent on because the image attached to the message

triggers an emotion in the sender, or a desire to show an

endorsement of that image.

58 We use this term to differentiate this from the very well-known media

psychology term selective exposure, where citizens favour information

which reinforces their pre-existing views. See Klapper, Joseph T. The effects

of mass communication . (Free Press, 1960) - (and a library of research after

that)

59 Of all the people reading this, who are also active on any social media,

90% of them have shared something without consuming. The other 10% are

lying to themselves. (Estimates, obviously!)

Image 1: WhatsApp message forwarded with a Jai Shree

Ram note appended to show fondness for Lord Rama, even

though the sender did not intend to endorse the actual

content of the message.

30

Image 2: A fairly typical example of unopened messages in a

WhatsApp group

I.4.2 Preference for images60: Images (memes or otherwise) or

image heavy messages are overwhelmingly preferred for

consumption or to engage with. Long text forwards and videos

are generally not preferred: the time/ effort involved in

consuming these and the low storage available in most phones

(for videos), are big barriers to the consumption of these

formats.

Given the demands on peoples attention and cognition, the

fact that they would be using visual cues to decide whether or

not to engage with the content does make sense; while the

claim that visuals are processed 60,000 times faster than

words is an internet myth61, there is certainly evidence that

pictures especially those with a positive or negative affect

(crudely, ability to arouse an emotional response) - are

processed faster than words.62

This is, in fact, one of the central discoveries of this project.

The canonical example of fake news, for example, the story

60 It could be argued that this is so much same old-same old. The history of

communicating in India more through images than through words is long

and rich. See for example Jain, Kajri. Gods in the bazaar : The economies of

Indian calendar art. (Durham : Duke University Press, 2007) and Pinney,

Christopher. 'Photos of the Gods' : The printed image and political struggle in

India. (London: Reaktion, 2004)

61 Or should we say visuals are processed 60,000 times faster than words is

fake news? 62 See, for example, Jan De Houwer, Dirk Hermans. Differences in the

Affective Proccessing of Words and Pictures . Cognition and Emotion , 1994:

1-20

31

about the Pope endorsing Donald Trump, created by

Macedonian teenagers, and circulating all over social media,

does not seem to be that prevalent in WhatsApp feeds in

India63. To be clearer, stories as a collection of words on a

website, circulated in the form of the url, do not seem to be the

most prevalent means of sharing information (and

disinformation).

The form of information that is consumed or engaged with

more, is visual information, sometimes layered with a

minimum amount of text.

Image 3: Translation: In a list of honest leaders issues in the

USA, there is only one person from Indian, that is Shree

Narendra Modi, and that too in first place (Note: Versions of

63 Note that this is based on observations of the respondents WhatsApp and

Facebook feeds. It would be useful for future researchers to establish exact

proportions through quantitative surveys.

this with ex- Prime Minister Manmohan Singh replacing

Narendra Modi are found too in WhatsApp/Facebook feeds).

Image 4: Translation Modi ji had said he would increase

GDP. He meant he would increase [the price of] Gas, Diesel,

Petrol. The failure to understand was ours, whats his fault?

Since people prefer to consume information in visual heavy

formats - or at best, very short text formats - they also prefer

to share in these formats. A very common mode of sharing is

32

through screenshots. In fact the practice of screenshotting is

something that has become relatively automatic and

instinctive!

There was inspiring news after long time search news had come. There was a program in Delhi for Kerala flood relief. Supreme Court lawyer and judge they did the singing and did the charity. I took the screenshot and forwarded. That time people came to know that judge can sing. (Male, 56, Rajkot)

Screenshotting (and storing it in the photo gallery) is used as

a method of sending something at a time later than the time

when the material has been consumed. Screenshots also act as

aide memoires:

Whatever I wish to purchase in gadgets I save its screenshots on my phoneI have kept it for my

collection of all the things I liked and want to look in detail about them.

(Male, 20, Udaipur)

Finally, because networks are relatively large and personal

relevance of messages is hard for the sender to assess,

screenshots become a way for the sender to signal to the

network that the sender does not know if this is relevant, but

will be happy to send more information (for example, the full

video) if required. It is also a way to compensate for the

possible low memory of the recipients phone and

consequent irritation if the message turns out to be irrelevant

but hogs memory.

Because I am sending screenshot to other person and if he or she finds it useful then they will message me that you send more details about it. If I am sending in the group, some person might think that this is useful but cannot download it properly, then that person might get irritated.

M: So, you feel it should send a small page?

R: Yes. If a person asks for more information, then I will send that person more details about that information personally.

M: So, you feel people will get irritated if you have shared the entire article?

33

R: Yes, since I myself feel irritated. Suppose I am not giving RPSC exam and somebody writes to me that this is very important please download it, I downloaded it but that was no use for me, so I am losing my MB, my memory is full, then I get some irritation (Female, 25, Udaipur)

This preference for images, screenshots and minimal text has

three key consequences, important for our purposes:

a) in depth information or analyses conveyed in words finds it

difficult to flourish in this environment

b) the headline often is the full story

c) it is harder for both the consumers - and for researchers/

analysts - to get to the source/creators of the information.

For example, in an environment that is predominantly image

heavy, the very effective digital techniques such as those

suggested in A Field Guide to Fake News and other

Information Disorders,64 become more difficult to pull off (as

most of the recipes in the Field Guide start with urls, which

in the case of images, rarely exist).

I.4.3. Sender primacy: Given the overwhelming amount of

information and the consequent inability or disinclination to

subject each message in ones feed to rational critical analysis,

people use heuristics to decide which content to engage with

and even to assess what they find credible. As such,

64 See the various recipes suggested at https://fakenews.publicdatalab.org/.

consumption decisions are disproportionately influenced by

which individual has sent the message in the first place. If that

individual happens to be a person of influence or social

standing in the recipients online or real life network, or

perhaps even more importantly an ideological/political leader

chances are that the message will be consumed and even

shared. On the flip side, certain individuals are labelled as

pointless spammers and their messages are avoided.

As I told you, I respect [xxx] Bhai, he is knowledgeable and knows more than me. So whatever, he sends me, I follow it. He has taught me a lot of things, I didnt know (Male, 24, Varanasi)

I trust anything my Mamaji sends, he knows a lot about the world. There are my other uncles

who stay in our hometown, I instantly mistrust anything they send. I dont even open most of

their forwards (Male, 27, Mumbai)

(On a piece of fake news) My friend has sent this to me, why would he send anything fake to me? (Female, 24, Vijayawada)

https://fakenews.publicdatalab.org/

34

XXXX is a leader who talks facts and figures. People may sleep in his speech but he talks

numbers. That is why when he shares anything or talks I just pay attention to it

(Female, 23, Raipur)

Note here the ages of the individuals quoted: they are all young people. Our analysis suggests that sender primacy operates across all ages and demographics, and does not have that much to do with levels of formal education.

I.4.4. Source agnosticism: This reliance on the identity of the

sender to judge the content of the messages is related to what

we are calling source agnosticism. That is, people really dont

seem to worry too much about where the messages they are

receiving originated from, especially on WhatsApp. It is not

that people dont know that there are more credible and less

credible sources. Nor is it the case that they dont care about

consuming incorrect information. Its that on the digital

platforms, while contending with the flood of onrushing

information, they simply cannot be bothered. The credibility

of the sender is what gives legitimacy to the message. The

original source, if at all present in the message itself, is often

ignored or unnoticed in Facebook,65 or completely absent in

WhatsApp. This is also happening, as detailed earlier, in an

65 This is true even after consumption. A private study commissioned by the

BBC World Service has established that on Facebook as less as 30% of the

audience recognised the source of a news item, even after they had consumed

the news item.

environment where the scepticism about the motivations and

intentions of the media is widespread- so its not as if there is

someone established as an arbiter of truth anyway. And as we

will see later, for many types of messages, the question of

where the message originated from becomes immaterial, since

they are anyway not about truth or lies, they are about

something much deeper than that.

I.4.5. The nature of the forum: The forum or composition of the

WhatsApp group something has been sent in becomes a key

determinant of engagement with and sharing of a message. It

matters whether the message is a personal one (i.e. sent one to

one) or whether the message has been sent in a group. If a

group, it matters what type of group it is: is it a family group,

or is it a group from work, or is it, in fact, a group dedicated to

politics - formal or informal? People think very actively about

the composition of their WhatsApp groups and therefore what

should be shared in which group. The same message

considered appropriate to share in one group can be

considered completely inappropriate to share in another

group.

I. 4.6: The affective more than the cognitive66: Given the deluge

of digital information and the (unarticulated but certainly

experienced) pressure to process information quickly, what is

engaged with more often than not, is what triggers an emotion.

And the way citizens often learn about things is through an

66 We are using these terms here in the sense used in the field of education/

learning. See, for a quick overview:

https://thesecondprinciple.com/instructional-design/threedomainsoflearning/

https://thesecondprinciple.com/instructional-design/threedomainsoflearning/

35

affective style, not a cognitive one. Put in non-academic

language, digital sharing platforms seem to be operating more

in the domain of feeling than thinking, though once in a

while, there is some critical engagement based on whether or

not something is of interest or relevance. Very importantly, in

the Indian case, what is engaged with and then shared is what

resonates with their ideology, especially their socio-political

identity. We will come back to this point later.

I.5. The broken link between consumption and

sharing:

For the ordinary citizen, consumption is hard, and

critically- engaged consumption is even harder. However,

sharing is easy. Digital sharing platforms are built to make

sharing easy - and there is no requirement built into the

platforms that sharing be done only after the consumption of

content. This means that there are very few considerations

that arrest the process of sharing, other than propriety and

relevance to the group that it is being shared in.

There is little verification by the citizen of the accuracy or

authenticity of the content that is being shared today. There

are some categories that are considered to be important for all:

non-political or soft news (for example, health) , updates (for

example, ATMs are going to be closed for five days) or policy

news ( for example, taxes have risen in the latest budget).

Because these are seen as being of universal relevance,

considerations of relevance do not come into play and they

are shared very widely and very quickly.

Health news and information - I will share with

all, someone will benefit out of it

I.6 Sharing Strategies on WhatsApp and

Facebook

Our two platforms of focus, WhatsApp and Facebook,

are used quite differently and quite strategically by citizens,

but in their own ways ,both end up facilitating conditions

conducive to the sharing of fake news.

How WhatsApp is used by citizens

WhatsApp is today the indispensable67 platform for citizens. It

is also the platform de jure to share. It is the platform to

connect with people you well - have a connection with. The

connection could be personal (ie family/ friends), or slightly

less personal (larger township, where it could include

strangers), or it could be a group about politics (a group for a

RSS shakha, say). But every group in WhatsApp has

something that bonds the group together - and makes it

behave like a group. That is, WhatsApp groups have group

norms, and it is seen as socially quite problematic to breach

those group norms. Generally speaking, these groups once

67 Strictly speaking we cant make this claim from a qualitative study; any

quantitative study should be able to easily validate, this, though.

36

created - do not grow too fast, and because there are definite

criteria for membership of groups, there is a natural limit to

their size.

Very importantly for our purposes, sharing in WhatsApp

groups is very targeted. WhatsApp usage is about validation of

ones beliefs and identities through the sharing of news and

information. Messages that originate in one group dont find it

that easy to move to another group. People are acutely

conscious of which messages belong in which groups. The

most actively political of persons might post a rabidly anti

Congress or BJP message in the political group s/he is part of,

but s/he will not post that same message in a family group.

I have not forwarded this to my lawyers group, I am in the minority there, they are all different kinds of people there. I share things in my cousins group, many like-minded people there, we can have discussions (Male, 62, Chennai)

However, if that person is part of multiple political groups, she

will share the same message in all of the political groups. This,

though, imposes costs in terms of time and effort. There isnt

any easy one click way of sharing with everyone in every

group you are part of, thereby imposing a threshold on the

number of people any message can be shared with relative

ease.

Image 4: The same political message about deshdrohi or

traitor to the nation- shared in multiple political groups

(some groups masked to preserve respondent anonymity).

Notethe preponderance of the saffron colour, which

indicates this person is a strongly right wing Hindu

nationalist.68

68 Note also the beer mug emojis in the name of the group at the bottom, right

next to the saffron flag, quite delightful touch!

37

This suggests that the chances of a fake news message

spreading on a nationwide scale on WhatsApp might actually

be quite limited. The defining feature of WhatsApp groups in

India might then not be its reach or scale or speed of

transmission of messages, but the fact that it is enabling

homophily,69 or the drawing together of people in tight

networks of like-mindedness.

But because of this this tightness, then, we suggest that it is

possible to use WhatsApp to mobilise. This starts to explain

why WhatsApp has seemed quite central to some cases of

violence in India. Its not the speed, or the reach of WhatsApp

that has been central to these issues, but the homophily of its

groups that has enabled mobilization in the cause of violence.

The platform is also very much known for spreading low

quality information of doubtful veracity. Most citizens are

aware of and have personally encountered messages that are

either trying to scam them, or peddle information that is

considered quite fantastical.

Despite that, WhatsApp is also a high engagement platform in

India with a high usage for news and news-related discussion-

and, concomitantly, a high amount of fake news. There is

certainly deliberate sharing of material here to start

69 See Miller McPherson, Lynn Smith-Lovin, James M Cook. Birds of a

Feather: Homophily in Social Networks. Annual Review of Sociology, 2001:

415-444

conversations or contribute to existing discussions, even if

these discussions rarely get full blown.

At the same time, the range of discussions, or the ability for

opposing viewpoints to find an airing is limited. There might

not be an algorithm like in the case of Facebook, but many of

the groups operate under unsaid rules about sharing what

everyone in the group approves of, or not sharing at all. There

are also some said agreements about what to share and not to

share in particular types of groups, with admins of the groups

playing an active role in weeding out those who breach group

rules, and individual members too policing content they feel

does not belong in that group, given the composition or the

stated intent behind the setting up of that group.

I am from the Kanyakubj samaj, we have like-minded people, with similar beliefs, so we share many things on that without thinking. I dont do it on Raipur Doctors group (Male, 34, Raipur)

M: Do you discuss only cricket related topics here?

R: Yes. It tells us about the trip series match, timings, the trophy picture is put up. I have a few friends in this group who are in favour of Modi and a few who are against him they constantly

38

have fights over this. I just comment once or twice here. If it goes on then the Admin of the group tries to calm them saying this is a cricket group, no Kejriwal or Modi discussions. At times their arguments heat up so much that till 1am the phone keeps buzzing. I have to mute the group. I have not made a group to discuss politics specially (Male, 41, Delhi)

Image 5: Response to a posted message: Please do not share

such messages. Group admin please pay attention to this

message

39

How Facebook is used by citizens:

Facebook, however, is very clearly perceived as a broadcasting

platform, with the main source of news being the newsfeed,

rather than one on one shares. Citizens do actively curate

their Facebook feed, liking the pages of various news or news

like organizations. Facebook is not seen as a platform for

likeminded or hemophilic groups, especially since it

constitutes of many contacts who are added with very little

personal connection to them. This nature of Facebook- added

to the extreme ease of sharing with every one of your friends-

makes it a broadcastin